Nemotron-3-120B stops responding mid-generation | Returns to "Idle" status

[andresl123]

When I'm interacting with the agent and ask it to do something, after some time, it just stops in the middle of generating and shows the status as 'connected | idle.' I'm using inference/nvidia/nemotron-3-super-120b-a12b on the cloud. Has anyone else faced this problem?

I'm using the latest version of OpenClaw. I created a custom image from the official NVIDIA image by changing the OpenClaw version (it could be a problem?).

[andresl123]

After updating Nemoclaw and Openclaw, I didn't see the same behavior as described previously.

[nv-cyuen]

Hi andresl123,

I'm glad it's working for you now! I'm going to close this discussion now.